Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

RECAP: Resistance Capture in Text-based Mental Health Counseling with Large Language Models

Anqi Li, Yuqian Chen, Yu Lu, Zhaoming Chen, Yuan Xie, Zhenzhong Lan

TL;DR

This work tackles the challenge of detecting client resistance in text-based mental health counseling by introducing PsyFIRE, a theoretically grounded taxonomy of 13 fine-grained resistance behaviors alongside collaboration. Building on PsyFIRE, the authors construct the ClientResistance corpus (23,930 annotated utterances from real-world Mandarin counseling) and develop RECAP, a two-stage, explainable framework that detects resistance and its subtypes with contextual rationales. RECAP outperforms strong LLM baselines, achieving 91.25% F1 in binary resistance detection and 66.58% macro-F1 in fine-grained classification, with significant gains when rationale generation is included. Additional analyses on an independent counseling dataset and a proof-of-concept study with 62 counselors demonstrate the model's prevalence insights, its relationship with therapeutic alliance, and its potential to improve counselor interventions through model-based feedback. The work provides a foundation for scalable, interpretable analysis of resistance in text-based therapy, with implications for training and real-time guidance, while acknowledging limitations related to cross-cultural generalizability and ethical deployment.

Abstract

Recognizing and navigating client resistance is critical for effective mental health counseling, yet detecting such behaviors is particularly challenging in text-based interactions. Existing NLP approaches oversimplify resistance categories, ignore the sequential dynamics of therapeutic interventions, and offer limited interpretability. To address these limitations, we propose PsyFIRE, a theoretically grounded framework capturing 13 fine-grained resistance behaviors alongside collaborative interactions. Based on PsyFIRE, we construct the ClientResistance corpus with 23,930 annotated utterances from real-world Chinese text-based counseling, each supported by context-specific rationales. Leveraging this dataset, we develop RECAP, a two-stage framework that detects resistance and fine-grained resistance types with explanations. RECAP achieves 91.25% F1 for distinguishing collaboration and resistance and 66.58% macro-F1 for fine-grained resistance categories classification, outperforming leading prompt-based LLM baselines by over 20 points. Applied to a separate counseling dataset and a pilot study with 62 counselors, RECAP reveals the prevalence of resistance, its negative impact on therapeutic relationships and demonstrates its potential to improve counselors' understanding and intervention strategies.

RECAP: Resistance Capture in Text-based Mental Health Counseling with Large Language Models

TL;DR

This work tackles the challenge of detecting client resistance in text-based mental health counseling by introducing PsyFIRE, a theoretically grounded taxonomy of 13 fine-grained resistance behaviors alongside collaboration. Building on PsyFIRE, the authors construct the ClientResistance corpus (23,930 annotated utterances from real-world Mandarin counseling) and develop RECAP, a two-stage, explainable framework that detects resistance and its subtypes with contextual rationales. RECAP outperforms strong LLM baselines, achieving 91.25% F1 in binary resistance detection and 66.58% macro-F1 in fine-grained classification, with significant gains when rationale generation is included. Additional analyses on an independent counseling dataset and a proof-of-concept study with 62 counselors demonstrate the model's prevalence insights, its relationship with therapeutic alliance, and its potential to improve counselor interventions through model-based feedback. The work provides a foundation for scalable, interpretable analysis of resistance in text-based therapy, with implications for training and real-time guidance, while acknowledging limitations related to cross-cultural generalizability and ethical deployment.

Abstract

Recognizing and navigating client resistance is critical for effective mental health counseling, yet detecting such behaviors is particularly challenging in text-based interactions. Existing NLP approaches oversimplify resistance categories, ignore the sequential dynamics of therapeutic interventions, and offer limited interpretability. To address these limitations, we propose PsyFIRE, a theoretically grounded framework capturing 13 fine-grained resistance behaviors alongside collaborative interactions. Based on PsyFIRE, we construct the ClientResistance corpus with 23,930 annotated utterances from real-world Chinese text-based counseling, each supported by context-specific rationales. Leveraging this dataset, we develop RECAP, a two-stage framework that detects resistance and fine-grained resistance types with explanations. RECAP achieves 91.25% F1 for distinguishing collaboration and resistance and 66.58% macro-F1 for fine-grained resistance categories classification, outperforming leading prompt-based LLM baselines by over 20 points. Applied to a separate counseling dataset and a pilot study with 62 counselors, RECAP reveals the prevalence of resistance, its negative impact on therapeutic relationships and demonstrates its potential to improve counselors' understanding and intervention strategies.
Paper Structure (51 sections, 2 figures, 13 tables)

This paper contains 51 sections, 2 figures, 13 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. How to Measure Resistance?
  4. Computational Approaches for Resistance Detection.
  5. Resistance Taxonomy
  6. Development Process
  7. Taxonomy
  8. Data Collection
  9. Data Source
  10. Annotation Process
  11. RECAP: Model for Resistance Capture
  12. Problem Definition
  13. Experimental Setup
  14. Baselines
  15. Experiment Results
  16. ...and 36 more sections

Figures (2)

  • Figure 1: A counseling dyad example showing that identical counselor interventions can elicit different client resistance behaviors, each requiring tailored counselor responses.
  • Figure 2: Left: Donut charts showing the distribution of collaboration versus resistance and fine-grained resistance types in the CounselingWAI dataset, as predicted by RECAP. Right: Pre- and post-test mean scores (±SD) for control and experimental groups, with ANOVA results from the proof-of-concept study.